CN117010554A - Dynamic multi-objective optimization method and device applied to e-commerce recommendation system - Google Patents
Dynamic multi-objective optimization method and device applied to e-commerce recommendation system Download PDFInfo
- Publication number
- CN117010554A CN117010554A CN202310806814.3A CN202310806814A CN117010554A CN 117010554 A CN117010554 A CN 117010554A CN 202310806814 A CN202310806814 A CN 202310806814A CN 117010554 A CN117010554 A CN 117010554A
- Authority
- CN
- China
- Prior art keywords
- commerce recommendation
- commerce
- constraint
- network
- objective
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000006870 function Effects 0.000 claims description 59
- 238000013528 artificial neural network Methods 0.000 claims description 21
- 238000013461 design Methods 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 11
- 238000012512 characterization method Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 description 11
- 238000004891 communication Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]
Abstract
The embodiment of the disclosure provides a dynamic multi-objective optimization method and device applied to an e-commerce recommendation system. The method comprises the following steps: acquiring an electronic commerce recommendation constraint condition and an electronic commerce recommendation objective function of an electronic commerce recommendation system; characterizing an electronic commerce recommendation constraint condition of the electronic commerce recommendation system as an electronic commerce recommendation constraint network, and characterizing an electronic commerce recommendation objective function of the electronic commerce recommendation system as an electronic commerce recommendation objective network; connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator; dynamic multi-objective optimization of the E-commerce recommendation system is performed based on the simulator, so that limited computing resources can be utilized, and the dynamic multi-objective optimization efficiency of the E-commerce recommendation system is improved.
Description
Technical Field
The disclosure relates to the technical field of computers, in particular to a dynamic multi-objective optimization method and device applied to an e-commerce recommendation system.
Background
In the e-commerce recommendation system, there is a Multi-objective optimization problem (Multi-objective Optimization Problem, MOP), and this type of problem has a plurality of objective functions that need to be optimized, and meanwhile, the objective functions are mutually restricted, so that the promotion of one objective function may cause the deterioration of other objective functions. Among the many Multi-objective optimization problems, there is a special Multi-objective optimization problem, namely Dynamic Multi-objective optimization problem (DMOP), which not only restricts objective functions to each other, but also changes the problem with time, and in addition, the dimension of data is very high, which results in that the traditional optimization scheme needs to occupy more computing resources and optimization time, and the optimization efficiency is low. Therefore, how to improve the dynamic multi-objective optimization efficiency of the e-commerce recommendation system becomes a technical problem to be solved at present.
Disclosure of Invention
The embodiment of the disclosure provides a dynamic multi-objective optimization method and device applied to an e-commerce recommendation system.
In a first aspect, an embodiment of the present disclosure provides a dynamic multi-objective optimization method applied to an e-commerce recommendation system, the method including:
acquiring an electronic commerce recommendation constraint condition and an electronic commerce recommendation objective function of an electronic commerce recommendation system;
characterizing an electronic commerce recommendation constraint condition of the electronic commerce recommendation system as an electronic commerce recommendation constraint network, and characterizing an electronic commerce recommendation objective function of the electronic commerce recommendation system as an electronic commerce recommendation objective network;
connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator;
and carrying out dynamic multi-objective optimization of the E-commerce recommendation system based on the simulator.
In some implementations of the first aspect, obtaining an e-commerce recommendation constraint condition and an e-commerce recommendation objective function of the e-commerce recommendation system includes:
acquiring a design scene of an electronic commerce recommendation system;
and analyzing based on the design scene of the E-commerce recommendation system to obtain the E-commerce recommendation constraint condition and the E-commerce recommendation objective function.
In some implementations of the first aspect, characterizing the e-commerce recommendation constraint of the e-commerce recommendation system as an e-commerce recommendation constraint network includes:
sampling is carried out at the periphery of the constraint boundary of the decision variable value taking space to obtain two types of sample points, wherein one type of sample points are negative samples meeting the recommendation constraint conditions of the E-commerce, and the other type of sample points are positive samples not meeting the recommendation constraint conditions of the E-commerce;
training the deep neural network based on the positive sample and the negative sample to obtain a neural network fitting model of the constraint space, and recording the neural network fitting model as an electronic commerce recommended constraint network.
In some implementations of the first aspect, the algorithm employed for sampling is a latin hypercube sampling algorithm.
In some implementations of the first aspect, characterizing an e-commerce recommendation objective function of the e-commerce recommendation system as an e-commerce recommendation objective network includes:
and constructing a network description form of the decision variable according to the E-commerce recommendation target function, and recording the network description form as an E-commerce recommendation target network.
In some implementations of the first aspect, the dynamic multi-objective optimization of the e-commerce recommendation system based on the simulator includes:
randomly generating a first generation individual set, inputting decision variables of the first generation individual to a simulator to obtain corresponding performance indexes and constraint losses, based on the corresponding performance indexes and constraint losses, obtaining dominant relations in a feasible solution set and an infeasible solution set based on a rapid non-dominant sorting algorithm, and selecting a high-quality individual set from non-dominant fronts of the feasible solution set and the infeasible solution set by using a reachability algorithm;
aiming at the high-quality individual set, analyzing the gradient direction of the decision variable for the performance index and the constraint loss by adopting an error back-pass algorithm, generating a next-generation individual set based on a gradient descent algorithm, iterating continuously until the iteration number exceeds the preset iteration number, stopping iterating, and taking the latest high-quality individual set as a dynamic multi-objective optimization result;
during the iteration, if the current objective function is detected to be changed, updating the E-commerce recommended objective network in the simulator according to the changed objective function, otherwise, continuing to use the existing simulator.
In some implementations of the first aspect, the method further includes:
and carrying out optimization adjustment on the E-commerce recommendation system according to the dynamic multi-objective optimization result.
In a second aspect, embodiments of the present disclosure provide a dynamic multi-objective optimization apparatus applied to an e-commerce recommendation system, the apparatus comprising:
the acquiring module is used for acquiring the E-commerce recommendation constraint conditions and the E-commerce recommendation objective function of the E-commerce recommendation system;
the characterization module is used for characterizing the E-commerce recommendation constraint condition of the E-commerce recommendation system as an E-commerce recommendation constraint network and characterizing the E-commerce recommendation objective function of the E-commerce recommendation system as an E-commerce recommendation objective network;
the parallel module is used for connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator;
and the optimizing module is used for carrying out dynamic multi-objective optimization of the E-commerce recommending system based on the simulator.
In a third aspect, embodiments of the present disclosure provide an electronic device comprising: at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method as described above.
In a fourth aspect, embodiments of the present disclosure provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform a method as described above.
In the embodiment of the disclosure, the electronic commerce recommendation constraint condition and the electronic commerce recommendation objective function of the electronic commerce recommendation system can be obtained; characterizing an electronic commerce recommendation constraint condition of the electronic commerce recommendation system as an electronic commerce recommendation constraint network, and characterizing an electronic commerce recommendation objective function of the electronic commerce recommendation system as an electronic commerce recommendation objective network; connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator; dynamic multi-objective optimization of the E-commerce recommendation system is performed based on the simulator, so that limited computing resources can be utilized, and the dynamic multi-objective optimization efficiency of the E-commerce recommendation system is improved.
It should be understood that what is described in this summary is not intended to limit the critical or essential features of the embodiments of the disclosure nor to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following description.
Drawings
The above and other features, advantages and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. For a better understanding of the present disclosure, and without limiting the disclosure thereto, the same or similar reference numerals denote the same or similar elements, wherein:
FIG. 1 illustrates a flow chart of a dynamic multi-objective optimization method applied to an e-commerce recommendation system provided by an embodiment of the present disclosure;
FIG. 2 is a block diagram of a dynamic multi-objective optimization device for an e-commerce recommendation system according to an embodiment of the present disclosure;
fig. 3 illustrates a block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present disclosure more apparent, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure, and it is apparent that the described embodiments are some embodiments of the present disclosure, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the disclosure, are within the scope of the disclosure.
In addition, the term "and/or" herein is merely an association relationship describing an association object, and means that three relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist together, and B exists alone. In addition, the character "/" herein generally indicates that the front and rear associated objects are an "or" relationship.
Aiming at the problems in the background art, the embodiment of the disclosure provides a dynamic multi-objective optimization method and device applied to an e-commerce recommendation system. Specifically, an electronic commerce recommendation constraint condition and an electronic commerce recommendation objective function of an electronic commerce recommendation system can be obtained; characterizing an electronic commerce recommendation constraint condition of the electronic commerce recommendation system as an electronic commerce recommendation constraint network, and characterizing an electronic commerce recommendation objective function of the electronic commerce recommendation system as an electronic commerce recommendation objective network; connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator; dynamic multi-objective optimization of the E-commerce recommendation system is performed based on the simulator, so that limited computing resources can be utilized, and the dynamic multi-objective optimization efficiency of the E-commerce recommendation system is improved.
The terms involved in the embodiments of the present disclosure are explained first below:
dynamic multi-objective optimization: the minimized dynamic multi-objective optimization problem involved in embodiments of the present disclosure is defined as follows:
minf(x,t)＝{f 1 (x,t),f 2 (x,t),…,f m (x,t)}(1)
s.t.g i (x)≥0h j (x)＝0i＝1,2,…,p,j＝1,2,…,q(2)
wherein t is an environmental (time) variable and x is R n The n-dimensional decision vector on the table, f (x, t) is an objective function, m is the number of the objective functions, and g (x, t) and h (x, t) are the corresponding inequality constraint and equality constraint. In most scenarios, the constraints are often not time-varying, and the objective function, i.e. the optimization objective, is dynamically changing, so this constraint time-invariant dynamic multi-objective optimization problem is discussed here with emphasis.
Pareto governs: at time t, for any two individuals x in the population 1 And x 2 If x 1 And x 2 The following conditions are satisfied:
then consider x 1 Dominant x 2 Is recorded as
Pareto optimal solution: at time t, there is no individual subject to individual x ′ ∈R n Then x is a Pareto optimal solution of the problem at time. At time t, all Pareto optimal solutions of the problem form PS t Expressed as:
pareto optimal leading edge, PF: at time t, S t Mapping in target space is called PF t Denoted as PF t ＝{F(x,t)|x∈PS t }。
The dynamic multi-objective optimization method and device applied to the e-commerce recommendation system provided by the embodiment of the disclosure are described in detail through specific embodiments with reference to the accompanying drawings.
Fig. 1 shows a flowchart of a dynamic multi-objective optimization method applied to an e-commerce recommendation system according to an embodiment of the present disclosure, and as shown in fig. 1, the dynamic multi-objective optimization method 100 may include the following steps:
s110, acquiring an electronic commerce recommendation constraint condition and an electronic commerce recommendation objective function of the electronic commerce recommendation system.
In some embodiments, a design scene of the e-commerce recommendation system may be obtained, and automatic and/or manual analysis may be performed based on the design scene of the e-commerce recommendation system, so as to quickly obtain e-commerce recommendation constraint conditions and e-commerce recommendation objective functions.
S120, characterizing the E-commerce recommendation constraint condition of the E-commerce recommendation system as an E-commerce recommendation constraint network, and characterizing the E-commerce recommendation objective function of the E-commerce recommendation system as an E-commerce recommendation objective network.
In some embodiments, the Latin hypercube sampling algorithm can be used for sampling around the constraint boundary of the decision variable value space to obtain two types of sample points, wherein one type of sample points is a negative sample meeting the E-commerce recommendation constraint condition, the other type of sample points is a positive sample not meeting the E-commerce recommendation constraint condition, the deep neural network is trained based on the positive sample and the negative sample, a neural network fitting model of the constraint space is obtained, and the neural network fitting model is recorded as the E-commerce recommendation constraint network. Wherein, the E-commerce recommended constraint network is kept unchanged in the subsequent execution process.
The above steps essentially map constraint boundaries into a two-class hyperplane using the universal fit characteristics of the deep neural network.
In other embodiments, the network description form of the decision variable may be constructed according to the e-commerce recommendation objective function, for example, the e-commerce recommendation objective function is represented by using a multi-objective neural network and is recorded as the e-commerce recommendation objective network. The e-commerce recommendation target network is updated along with the change of the e-commerce recommendation target function in the subsequent execution process.
The above steps are essentially direct translation processes, and the parameters of the target network are parameters of the target function, and training is not required.
S130, connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form the simulator.
For convenience of processing, the output of the simulator is minimized, and the output of the constraint network can be regarded as constraint loss, namely the degree to which an individual does not meet constraint conditions, and the smaller the value, the better the e-commerce recommends that the output of the constraint network be a performance index.
S140, dynamic multi-objective optimization of the E-commerce recommendation system is performed based on the simulator.
Specifically, a first generation individual set is randomly generated, decision variables of the first generation individuals are input into a simulator to obtain corresponding performance indexes and constraint losses, on the basis of the corresponding performance indexes and constraint losses, the dominant relation in a feasible solution set and an infeasible solution set is obtained based on a rapid non-dominant sorting algorithm, and a reachability algorithm is used for selecting a high-quality individual set from the non-dominant fronts of the feasible solution set and the infeasible solution set.
And analyzing the gradient direction of the decision variable to the performance index and the constraint loss by adopting an error back-pass algorithm aiming at the high-quality individual set, generating a next-generation individual set based on a gradient descent algorithm, iterating continuously until the iteration number exceeds the preset iteration number, stopping iterating, and taking the latest high-quality individual set as a dynamic multi-target optimization result.
Illustratively, generating the next generation set of individuals based on the gradient descent algorithm may be as follows:
let the ith performance indicator be denoted as f i Constraint loss is noted as g, and the following gradient descent algorithm obtains the next generation set of individuals:
wherein x is n+1 Individual decision variables, x, representing the next generation n Representing the current generation of individual decision variables, alpha, beta being the learning rate.
It is noted that during the iteration, if the current objective function is detected to change, the e-commerce recommended objective network in the simulator is updated according to the changed objective function, otherwise, the existing simulator is continuously used.
In summary, the embodiments of the present disclosure bring the following technical effects:
1. and the constraint conditions are fitted by using a neural network, so that the high-dimensional data can be processed by using the neural network data, the consumption of computing resources is reduced, and the computing efficiency is improved.
2. The objective function, namely the optimization objective, is directly constructed as a neural network and is connected in parallel with the constraint network to form the simulator. The output of the simulator comprises constraint loss of the individual to the constraint condition and performance index of the target function, and when the target function is changed, the new scene can be adapted only by adjusting the structure of the target network.
3. By utilizing the error back-propagation algorithm, new and better individuals are more efficiently found, and the method is faster to descend compared with the traditional evolutionary or other intelligent algorithms.
4. When the objective function is changed, namely the environment is changed, the constraint conditions are unchanged, so that multiplexing can be realized, and the optimization efficiency is improved.
5. For optimization problems with special requirements (such as preference among targets, strict constraint conditions, etc.), the optimization problem can be realized by directly adjusting the weights of the constraint network and the target network manually.
Notably, in order to improve the recommendation effect of the e-commerce recommendation system, the dynamic multi-objective optimization method 100 may further include: and generating an adjustment plan according to the dynamic multi-objective optimization result, and carrying out optimization adjustment on the E-commerce recommendation system.
The dynamic multi-objective optimization method 100 provided by the embodiments of the present disclosure is described in detail below with reference to a specific embodiment, which is specifically as follows:
in the design scene of the e-commerce recommendation system, market demands, business targets and distribution constraint conditions are always changed continuously, and how to optimize various experience indexes of users of the e-commerce recommendation system in the flow distribution process and maximize the long-term benefits of a platform is a typical dynamic multi-target optimization problem. In this regard, the dynamic multi-objective optimization method 100 may include:
(1) And acquiring a design scene of the E-commerce recommendation system, and automatically and/or manually analyzing the design scene of the E-commerce recommendation system, so as to quickly acquire the E-commerce recommendation constraint condition and the E-commerce recommendation objective function.
(2) Based on Latin hypercube sampling algorithm, sampling is carried out around the constraint boundary of the decision variable value space to obtain positive and negative samples, the deep neural network is trained based on the positive and negative samples, the neural network fitting model of the constraint space is obtained, and the neural network fitting model is recorded as the E-commerce recommended constraint network.
(3) And the E-commerce recommendation target function is expressed by using a multi-target neural network and is recorded as the E-commerce recommendation target network.
(4) And connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form the simulator.
(5) And randomly generating a first generation individual set, inputting decision variables of the first generation individuals into a simulator to obtain corresponding performance indexes and constraint losses, based on the corresponding performance indexes and constraint losses, obtaining dominant relations in a feasible solution set and an infeasible solution set based on a rapid non-dominant sorting algorithm, and selecting a high-quality individual set from non-dominant fronts of the feasible solution set and the infeasible solution set by using a reachability algorithm.
Judging whether the current iteration times exceeds the preset iteration times, stopping iteration, taking the latest high-quality individual set as a dynamic multi-objective optimization result if the current iteration times exceed the preset iteration times, otherwise detecting whether the current objective function is changed, updating an e-commerce recommended objective network in the simulator according to the changed objective function if the current objective function is changed, otherwise continuously using the existing simulator, analyzing the gradient directions of decision variables for performance indexes and constraint losses by adopting an error back-propagation algorithm for the latest high-quality individual set, and generating a next generation individual set based on a gradient descent algorithm for continuous iteration.
(6) And generating an adjustment plan according to the dynamic multi-objective optimization result, and carrying out optimization adjustment on the E-commerce recommendation system.
Therefore, the dynamic multi-objective optimization problem can be solved by utilizing limited computing resources under the complex condition that the decision variable dimension is high and the objective is time-varying.
It should be noted that, for simplicity of description, the foregoing method embodiments are all described as a series of acts, but it should be understood by those skilled in the art that the present disclosure is not limited by the order of acts described, as some steps may be performed in other orders or concurrently in accordance with the present disclosure. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all alternative embodiments, and that the acts and modules referred to are not necessarily required by the present disclosure.
The foregoing is a description of embodiments of the method, and the following further describes embodiments of the present disclosure through examples of apparatus.
Fig. 2 illustrates a block diagram of a dynamic multi-objective optimization apparatus applied to an e-commerce recommendation system according to an embodiment of the present disclosure, and as shown in fig. 2, the dynamic multi-objective optimization apparatus 200 may include:
the acquiring module 210 is configured to acquire an e-commerce recommendation constraint condition and an e-commerce recommendation objective function of the e-commerce recommendation system.
The characterization module 220 is configured to characterize the e-commerce recommendation constraint condition of the e-commerce recommendation system as an e-commerce recommendation constraint network, and characterize the e-commerce recommendation objective function of the e-commerce recommendation system as an e-commerce recommendation objective network.
The parallel module 230 is configured to connect the e-commerce recommendation constraint network and the e-commerce recommendation target network in parallel to form a simulator.
The optimizing module 240 is configured to perform dynamic multi-objective optimization of the e-commerce recommendation system based on the simulator.
In some embodiments, the obtaining module 210 is specifically configured to:
acquiring a design scene of an electronic commerce recommendation system;
and analyzing based on the design scene of the E-commerce recommendation system to obtain the E-commerce recommendation constraint condition and the E-commerce recommendation objective function.
In some embodiments, characterization module 220 is specifically configured to:
sampling is carried out at the periphery of the constraint boundary of the decision variable value taking space to obtain two types of sample points, wherein one type of sample points are negative samples meeting the recommendation constraint conditions of the E-commerce, and the other type of sample points are positive samples not meeting the recommendation constraint conditions of the E-commerce;
training the deep neural network based on the positive sample and the negative sample to obtain a neural network fitting model of the constraint space, and recording the neural network fitting model as an electronic commerce recommended constraint network.
In some embodiments, characterization module 220 is specifically configured to:
and constructing a network description form of the decision variable according to the E-commerce recommendation target function, and recording the network description form as an E-commerce recommendation target network.
In some embodiments, the optimization module 240 is specifically configured to:
randomly generating a first generation individual set, inputting decision variables of the first generation individual to a simulator to obtain corresponding performance indexes and constraint losses, based on the corresponding performance indexes and constraint losses, obtaining dominant relations in a feasible solution set and an infeasible solution set based on a rapid non-dominant sorting algorithm, and selecting a high-quality individual set from non-dominant fronts of the feasible solution set and the infeasible solution set by using a reachability algorithm;
aiming at the high-quality individual set, analyzing the gradient direction of the decision variable for the performance index and the constraint loss by adopting an error back-pass algorithm, generating a next-generation individual set based on a gradient descent algorithm, iterating continuously until the iteration number exceeds the preset iteration number, stopping iterating, and taking the latest high-quality individual set as a dynamic multi-objective optimization result;
during the iteration, if the current objective function is detected to be changed, updating the E-commerce recommended objective network in the simulator according to the changed objective function, otherwise, continuing to use the existing simulator.
In some embodiments, the dynamic multi-objective optimization apparatus 200 further comprises:
and the adjustment module is used for carrying out optimization adjustment on the E-commerce recommendation system according to the dynamic multi-objective optimization result.
It can be appreciated that each module/unit in the dynamic multi-objective optimization apparatus 200 shown in fig. 2 has a function of implementing each step in the dynamic multi-objective optimization method 100 shown in fig. 1, and can achieve corresponding technical effects, which are not described herein for brevity.
Fig. 3 illustrates a block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure. Electronic device 300 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic device 300 may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.
As shown in fig. 3, the electronic device 300 may include a computing unit 301 that may perform various suitable actions and processes in accordance with a computer program stored in a Read Only Memory (ROM) 302 or a computer program loaded from a storage unit 308 into a Random Access Memory (RAM) 303. In the RAM303, various programs and data required for the operation of the electronic device 300 may also be stored. The computing unit 301, the ROM302, and the RAM303 are connected to each other by a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.
Various components in the electronic device 300 are connected to the I/O interface 305, including: an input unit 306 such as a keyboard, a mouse, etc.; an output unit 307 such as various types of displays, speakers, and the like; a storage unit 308 such as a magnetic disk, an optical disk, or the like; and a communication unit 309 such as a network card, modem, wireless communication transceiver, etc. The communication unit 309 allows the electronic device 300 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.
The computing unit 301 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 301 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, the method 100 may be implemented as a computer program product, including a computer program, tangibly embodied on a computer-readable medium, such as the storage unit 308. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 300 via the ROM302 and/or the communication unit 309. One or more of the steps of the method 100 described above may be performed when the computer program is loaded into RAM303 and executed by the computing unit 301. Alternatively, in other embodiments, the computing unit 301 may be configured to perform the method 100 by any other suitable means (e.g. by means of firmware).
The various embodiments described above herein may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems-on-a-chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of this disclosure, a computer-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. The computer readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a computer-readable storage medium would include one or more wire-based electrical connections, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
It should be noted that the present disclosure further provides a non-transitory computer readable storage medium storing computer instructions, where the computer instructions are configured to cause a computer to perform the method 100 and achieve corresponding technical effects achieved by performing the method according to the embodiments of the present disclosure, which are not described herein for brevity.
In addition, the present disclosure also provides a computer program product comprising a computer program which, when executed by a processor, implements the method 100.
To provide for interaction with a user, the embodiments described above may be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The above-described embodiments may be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.
The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the disclosed aspects are achieved, and are not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.
Claims (10)
1. A dynamic multi-objective optimization method applied to an e-commerce recommendation system, the method comprising:
acquiring an electronic commerce recommendation constraint condition and an electronic commerce recommendation objective function of an electronic commerce recommendation system;
characterizing an electronic commerce recommendation constraint condition of the electronic commerce recommendation system as an electronic commerce recommendation constraint network, and characterizing an electronic commerce recommendation objective function of the electronic commerce recommendation system as an electronic commerce recommendation objective network;
connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator;
and carrying out dynamic multi-objective optimization of the E-commerce recommendation system based on the simulator.
2. The method of claim 1, wherein the obtaining the e-commerce recommendation constraint condition and the e-commerce recommendation objective function of the e-commerce recommendation system comprises:
acquiring a design scene of the E-commerce recommendation system;
and analyzing based on the design scene of the E-commerce recommendation system to obtain E-commerce recommendation constraint conditions and E-commerce recommendation objective functions.
3. The method of claim 1, wherein characterizing the e-commerce recommendation constraint of the e-commerce recommendation system as an e-commerce recommendation constraint network comprises:
sampling is carried out at the periphery of the constraint boundary of the decision variable value taking space to obtain two types of sample points, wherein one type of sample points are negative samples meeting the recommendation constraint conditions of the E-commerce, and the other type of sample points are positive samples not meeting the recommendation constraint conditions of the E-commerce;
training the deep neural network based on the positive sample and the negative sample to obtain a neural network fitting model of the constraint space, and recording the neural network fitting model as an electronic commerce recommended constraint network.
4. A method according to claim 3, characterized in that the algorithm used for sampling is the latin hypercube sampling algorithm.
5. The method of claim 1, wherein characterizing the e-commerce recommendation objective function of the e-commerce recommendation system as an e-commerce recommendation objective network comprises:
and constructing a network description form of the decision variable according to the E-commerce recommendation target function, and recording the network description form as an E-commerce recommendation target network.
6. The method of claim 1, wherein the performing, based on the simulator, dynamic multi-objective optimization of the e-commerce recommendation system comprises:
randomly generating a first generation individual set, inputting decision variables of the first generation individuals into the simulator to obtain corresponding performance indexes and constraint losses, based on the corresponding performance indexes and constraint losses, obtaining dominant relations in a feasible solution set and an infeasible solution set based on a rapid non-dominant sorting algorithm, and selecting a high-quality individual set from non-dominant fronts of the feasible solution set and the infeasible solution set by using a reachability algorithm;
aiming at the high-quality individual set, analyzing the gradient direction of the decision variable for the performance index and the constraint loss by adopting an error back-pass algorithm, generating a next-generation individual set based on a gradient descent algorithm, iterating continuously until the iteration number exceeds the preset iteration number, stopping iterating, and taking the latest high-quality individual set as a dynamic multi-objective optimization result;
during the iteration, if the current objective function is detected to be changed, updating the E-commerce recommended objective network in the simulator according to the changed objective function, otherwise, continuing to use the existing simulator.
7. The method of claim 6, wherein the method further comprises:
and carrying out optimization adjustment on the E-commerce recommendation system according to the dynamic multi-objective optimization result.
8. A dynamic multi-objective optimization device for an e-commerce recommendation system, the device comprising:
the acquiring module is used for acquiring the E-commerce recommendation constraint conditions and the E-commerce recommendation objective function of the E-commerce recommendation system;
the characterization module is used for characterizing the E-commerce recommendation constraint condition of the E-commerce recommendation system as an E-commerce recommendation constraint network and characterizing the E-commerce recommendation objective function of the E-commerce recommendation system as an E-commerce recommendation objective network;
the parallel module is used for connecting the E-commerce recommendation constraint network and the E-commerce recommendation target network in parallel to form a simulator;
and the optimizing module is used for carrying out dynamic multi-objective optimization of the E-commerce recommendation system based on the simulator.
9. An electronic device, the electronic device comprising:
at least one processor;
and a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.
10. A non-transitory computer readable storage medium storing computer instructions for causing a computer to perform the method of any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310806814.3A CN117010554A (en) | 2023-07-03 | 2023-07-03 | Dynamic multi-objective optimization method and device applied to e-commerce recommendation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310806814.3A CN117010554A (en) | 2023-07-03 | 2023-07-03 | Dynamic multi-objective optimization method and device applied to e-commerce recommendation system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117010554A true CN117010554A (en) | 2023-11-07 |
Family
ID=88566418
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310806814.3A Pending CN117010554A (en) | 2023-07-03 | 2023-07-03 | Dynamic multi-objective optimization method and device applied to e-commerce recommendation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117010554A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200104678A1 (en) * | 2018-09-27 | 2020-04-02 | Google Llc | Training optimizer neural networks |
| CN111563791A (en) * | 2020-03-31 | 2020-08-21 | 北京奇艺世纪科技有限公司 | Commodity dividing and recommending method and device, electronic equipment and storage medium |
| CN112561167A (en) * | 2020-12-17 | 2021-03-26 | 北京百度网讯科技有限公司 | Travel recommendation method and device, electronic equipment and storage medium |
| CN113269617A (en) * | 2021-06-07 | 2021-08-17 | 大连理工大学 | Large-scale multi-objective optimization commodity recommendation method based on image coding |
| CN113761347A (en) * | 2021-02-25 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Commodity recommendation method, commodity recommendation device, storage medium and commodity recommendation system |
| CN114329280A (en) * | 2021-12-31 | 2022-04-12 | 中国电信股份有限公司 | Method and device for resource recommendation, storage medium and electronic equipment |
| CN114547446A (en) * | 2022-02-15 | 2022-05-27 | 深圳闪回科技有限公司 | Order pushing method and system |
-
2023
- 2023-07-03 CN CN202310806814.3A patent/CN117010554A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200104678A1 (en) * | 2018-09-27 | 2020-04-02 | Google Llc | Training optimizer neural networks |
| CN111563791A (en) * | 2020-03-31 | 2020-08-21 | 北京奇艺世纪科技有限公司 | Commodity dividing and recommending method and device, electronic equipment and storage medium |
| CN112561167A (en) * | 2020-12-17 | 2021-03-26 | 北京百度网讯科技有限公司 | Travel recommendation method and device, electronic equipment and storage medium |
| CN113761347A (en) * | 2021-02-25 | 2021-12-07 | 北京沃东天骏信息技术有限公司 | Commodity recommendation method, commodity recommendation device, storage medium and commodity recommendation system |
| CN113269617A (en) * | 2021-06-07 | 2021-08-17 | 大连理工大学 | Large-scale multi-objective optimization commodity recommendation method based on image coding |
| CN114329280A (en) * | 2021-12-31 | 2022-04-12 | 中国电信股份有限公司 | Method and device for resource recommendation, storage medium and electronic equipment |
| CN114547446A (en) * | 2022-02-15 | 2022-05-27 | 深圳闪回科技有限公司 | Order pushing method and system |
Non-Patent Citations (1)
| Title |
|---|
| 厍向阳;蔡院强;董立红;: "新的基于多目标优化的推荐算法", 计算机应用, no. 01, 10 January 2015 (2015-01-10) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111311321B (en) | User consumption behavior prediction model training method, device, equipment and storage medium | |
| CN112560996B (en) | User portrait identification model training method, device, readable storage medium and product | |
| CN111460384B (en) | Policy evaluation method, device and equipment | |
| CN114065863B (en) | Federal learning method, apparatus, system, electronic device and storage medium | |
| CN112084366A (en) | Method, apparatus, device and storage medium for retrieving image | |
| CN114119989A (en) | Training method and device for image feature extraction model and electronic equipment | |
| CN112580732A (en) | Model training method, device, equipment, storage medium and program product | |
| CN112561081A (en) | Conversion method and device of deep learning model, electronic equipment and storage medium | |
| CN112506359A (en) | Method and device for providing candidate long sentences in input method and electronic equipment | |
| CN110046670B (en) | Feature vector dimension reduction method and device | |
| CN117010554A (en) | Dynamic multi-objective optimization method and device applied to e-commerce recommendation system | |
| CN113722593B (en) | Event data processing method, device, electronic equipment and medium | |
| CN115185606A (en) | Method, device, equipment and storage medium for obtaining service configuration parameters | |
| CN112231299B (en) | Method and device for dynamically adjusting feature library | |
| CN114217933A (en) | Multi-task scheduling method, device, equipment and storage medium | |
| CN111695036B (en) | Content recommendation method and device | |
| CN114707638A (en) | Model training method, model training device, object recognition method, object recognition device, object recognition medium and product | |
| CN113033219A (en) | Model training method and device, electronic equipment and computer readable storage medium | |
| CN114417822A (en) | Method, apparatus, device, medium and product for generating model interpretation information | |
| CN113112311A (en) | Method for training causal inference model, information prompting method and device | |
| CN112905885A (en) | Method, apparatus, device, medium, and program product for recommending resources to a user | |
| CN112528103A (en) | Method and device for recommending objects | |
| CN113761379B (en) | Commodity recommendation method and device, electronic equipment and medium | |
| CN114331379B (en) | Method for outputting task to be handled, model training method and device | |
| CN114066278B (en) | Method, apparatus, medium, and program product for evaluating article recall |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |